Surface Sensing for Paenibacillus sp.NAIST15-1 Flagellar Gene Expression onSolid Medium

Kazuo Kobayashi,a Yu Kanesaki,b Hirofumi Yoshikawab,c

Graduate School of Biological Sciences, Nara Institute of Science and Technology, Ikoma, Japana; NODAIGenome Research Center, Tokyo University of Agriculture, Tokyo, Japanb; Department of Bioscience, TokyoUniversity of Agriculture, Tokyo, Japanc

ABSTRACT A rhizosphere Gram-positive bacterial isolate, Paenibacillus sp. NAIST15-1,exhibits intriguing motility behavior on hard agar medium. Paenibacillus sp. shows in-creased transcription of flagellar genes and hyperflagellation when transferred from liq-uid to solid medium. Hyperflagellated cells form wandering colonies that are capable ofmoving around on the surface of medium containing �1.5% agar. Transposon mu-tagenesis was used to identify genes critical for motility. In addition to flagellar genes,this mutagenesis identified five nonflagellar structural genes that were important formotility. Of these, the disruption of degSU, wsfP, or PBN151_4312 resulted in a completeloss of flagellin synthesis. Analysis of flagellar gene promoter activity showed that eachmutation severely reduced flagellar gene transcription in a different manner. Flagellargene transcription was induced in liquid medium by the addition of a viscous agent, Fi-coll, or by disruption of flagellar stator genes, indicating that flagellar gene transcriptionwas induced in response to restriction of flagellar rotation. Overexpression of DegSU by-passed the requirement of flagellar rotation restriction for induction of flagellar genes.These results indicate that physical restriction of flagellar rotation by physical contactwith the surface of solid medium induces flagellar gene transcription through the activa-tion of DegSU. Further analysis revealed that the same mechanism was conserved in Ba-cillus subtilis. These results demonstrate that flagella act as mechanosensors to controlflagellar transcription in Gram-positive bacteria.

IMPORTANCE Many bacteria exist on living or nonliving surfaces in nature. Bacteriaexpress distinct behaviors, such as surface motility and biofilm formation, to adaptto surfaces. However, it remains largely unknown how bacteria sense the surfaces onwhich they sit and how they induce the genes needed for growth on a surface.Swarming motility is flagellum-dependent motility on a surface. The Gram-positivebacterium Paenibacillus sp. exhibits strong swarming motility ability and is capableof moving on 1.5% agar medium. In this study, we showed that the two-componentsystem DegSU was responsible for inducing flagellar genes in response to heavyloads on flagellar rotation in Paenibacillus sp. The same mechanism was conservedin a related species, B. subtilis, even though these two bacteria exhibit very differentmotility behaviors. This study shows that flagellum serves as a sensor for surfacecontact to induce flagellar gene transcription in these bacteria.

KEYWORDS DegSU, Paenibacillus, flagellar rotation, surface sensing, swarmingmotility, wandering colony

Bacteria are capable of moving on a variety of surfaces. Swarming motility, which isdefined as flagellum-dependent group motility on a solid surface, is observed in

diverse bacteria (1–3). Swarming motility is distinct from swimming motility, which isflagellum-dependent movement of planktonic cells grown in an aqueous environmentand requires differentiation into specialized cells (2, 3). Swarmer cells express a different

Received 13 March 2017 Accepted 19 May2017

Accepted manuscript posted online 26 May2017

Citation Kobayashi K, Kanesaki Y, Yoshikawa H.2017. Surface sensing for Paenibacillus sp.NAIST15-1 flagellar gene expression on solidmedium. Appl Environ Microbiol 83:e00585-17.https://doi.org/10.1128/AEM.00585-17.

Editor Harold L. Drake, University of Bayreuth

Copyright © 2017 American Society forMicrobiology. All Rights Reserved.

Address correspondence to Kazuo Kobayashi,kazuok@bs.naist.jp.

GENETICS AND MOLECULAR BIOLOGY

crossm

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 1Applied and Environmental Microbiology

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

set of genes from those expressed by planktonic cells and often exhibit hyperflagella-tion, cell elongation, the formation of multicellular rafts, and the secretion of wettingagents that attract water to the surface or reduce surface tension (2, 3).

The initiation of swarmer cell differentiation is thought to start when bacterial cellssense the surface on which they are located (2, 3). A mechanism for surface sensing wasfirst discovered in Vibrio parahaemolyticus (4–7). In this mechanism, the transcriptionfactor LafK is activated when rotation of a polar flagellum is impeded by physicalcontact with a surface or by a highly viscous environment. LafK then induces theexpression of lateral flagella, which produce the driving force for swarming motility.Restriction of flagellar rotation likewise induces swarmer cell differentiation in otherbacteria. In Proteus mirabilis, a flagellar hook-basal body protein, FliL, is implicated toplay a role in surface sensing (8, 9). In Escherichia coli, the number of stator proteinsbound to flagella increases as flagellar rotational load increases (10, 11). In theseGram-negative bacteria, flagella appear to serve as a sensor for surface contact signals.

Although swarming motility was previously analyzed in many Gram-negative bac-teria, the only Gram-positive bacterium whose swarming motility system has beenextensively studied is Bacillus subtilis (reference 12 and references therein). In B. subtilis,the cytoplasmic protein SwrA, which is only conserved in B. subtilis and its closestrelatives, plays a critical role in inducing swarmer cell differentiation (13, 14). SwrAfacilitates binding of DegU, the response regulator of the DegSU two-componentsystem, to the promoter of the fla-che flagellar operon, leading to increased flagellartranscription (15, 16). SwrA levels are restricted by LonA-dependent proteolysis in liquidmedia but not on solid media; however, the mechanisms underlying this differenceremain unknown (17). Restriction of flagellar rotation activates DegSU in B. subtilis, butthis is thought to repress motility and lead to biofilm formation and poly-�-glutamateproduction (17–21). Thus, the surface-sensing mechanisms underlying swarmer celldifferentiation in Gram-positive bacteria remain unresolved.

Paenibacillus is a genus of Gram-positive facultative aerobic spore-forming bacteria(22) found in a wide variety of environments, particularly those associated with plants(23–27). Paenibacillus sp. NAIST15-1 (here referred to as Paenibacillus sp.) was previouslyisolated from a plant rhizosphere (28). Unlike many other Paenibacillus bacteria, Paeni-bacillus sp. is amenable to genetic manipulation (28). Paenibacillus sp. exhibits intrigu-ing motility behavior on hard agar media (28). When inoculated on 1.5% agar medium,many small Paenibacillus sp. colonies formed at the inoculation site and then movedover the surface at a speed of 3.6 �m/s. As colony density increased, colony trafficceased, and each colony rotated on the spot to eventually form a static round colony.This motility behavior is driven by flagella, which are strongly induced when Paeniba-cillus sp. was grown on solid medium. Despite its unusual motility behavior, theorganization and composition of flagellar genes in Paenibacillus sp. resemble those ofBacillus subtilis, with the exception that Paenibacillus sp. has no SwrA homolog andexpresses a large extracellular protein, CmoA, which is required for colony movement(28). While Paenibacillus bacteria are related to B. subtilis, sufficient phylogeneticdistance is observed to warrant assignment of a distinct genus (22). Thus, Paenibacillussp. is a good model for motility analysis not only because of its intriguing motilitybehavior but also because of its similarity to B. subtilis. We believe that comparativeanalysis of these bacteria will provide new insights into their swarming mechanisms.

Here, transposon mutagenesis was used to identify genes involved in the motility ofPaenibacillus sp. Three genes were identified that were critical for flagellar genetranscription. As in B. subtilis, transcription of flagellar genes in Paenibacillus sp. wascontrolled by early and late promoters. The three genes each had different effects onthese promoter activities. In particular, DegSU positively regulated the early promotercontrolling the fla-che flagellar operon. Flagellar gene transcription was induced byphysical and genetic inhibition of flagellar rotation. Overexpression of DegSU led toconstitutive transcription of flagellar genes in liquid and solid media. We propose thatDegSU activates flagellar gene transcription in response to heavy loads on flagellarrotation in Paenibacillus sp., and that this mechanism is in part conserved in B. subtilis.

Kobayashi et al. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 2

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

RESULTS AND DISCUSSIONIdentification of genes involved in motility. Paenibacillus sp. NAIST15-1 (here

referred to as Paenibacillus sp.) exhibits unusual motility behavior on hard agar mediumand is able to move on medium containing �1.5% agar (28). A mini-Tn10 deliveryplasmid, pIC333 (29), was used for transposon mutagenesis to identify genes requiredfor motility in Paenibacillus sp. When grown on 2.25% agar medium, Paenibacillus sp.did not spread on the surface of the medium but formed distinct colonies that wereirregularly shaped rather than round, probably as a result of its robust motile activity.We hypothesized that motility-defective mutants might form round colonies on 2.25%agar medium. Based on this hypothesis, approximately 20,000 colonies from eightindependent libraries were screened for Tn10 insertion mutants that formed roundcolonies on 2.25% agar. Round colonies were then picked, and their motility was testedon 1.5% agar plates. Ninety motility-defective mutants were isolated using this proce-dure. The DNA regions flanking the mini-Tn10 insertions were amplified using aninverse PCR technique, and their DNA sequences were determined. We identified 62mini-Tn10 insertion sites (Table 1). Many motility-defective mutants contained mini-Tn10 insertions within the 30-kb putative fla-che operon, which is a counterpart of theB. subtilis fla-che flagellar operon. This observation was consistent with the previousfinding that Paenibacillus sp. motility is driven by flagella (28).

In addition to insertions in the fla-che operon, Tn10 insertions were identified innonflagellar structural genes degS, wsfP, and fliB and several unknown genes (Table 1).To test whether these nonflagellar structural genes were indeed required for motility,disruption mutants of these genes were constructed using a pMAD plasmid (30). Thegrowth rates of the disruption mutants in 2� yeast extract-tryptone (YT) liquid mediumwere indistinguishable from those of the wild-type strain. The motility ability of thedisruption mutants was then tested on 2� YT medium containing 0.3 to 1.5% agar. TheDegS-DegU two-component system is an ortholog of B. subtilis DegS-DegU, whichactivates transcription of the fla-che flagellar operon (31). Disruption of degSU com-pletely abolished the motility of Paenibacillus sp. on all agar media tested (Fig. 1). Thedeleted degSU region on the genome was restored using a pMAD derivative plasmidcarrying the wild-type degSU sequence (see Materials and Methods). The resulting

TABLE 1 Mini-Tn10 insertion mutants

Target geneGene length(bp)

Insertion sitesof mini-Tn10(bp from start)

No. ofmini-Tn10insertions

Motility(0.5% agar)

Motility(1.5% agar) Gene annotation

flgC 453 216 1 Defective Defective Flagellar basal-body rod protein FlgCfliG 1,008 995 4 Defective Defective Flagellar motor switch protein FliGfliJ 447 335 3 Defective Defective Flagellar export protein FliJylzI 225 53 2 Defective Defective Putative flagellar protein YlzIylxF 918 289 10 Defective Defective flaA locus 22.9-kDa proteinfliM 999 428 3 Defective Defective Flagellar motor switch protein FliMfliY 1,266 844 1 Defective Defective Flagellar motor switch phosphatase

FliYsigD 786 347 2 Defective Defective RNA polymerase sigma factor SigDPBN151_1926 1,407 409 2 Defective Defective Hypothetical protein (gene

downstream of sigD)PBN151_0898 987 965 8 Defective Defective Peptide methionine sulfoxide reductase

MsrA/MsrBwsfP 1,407 591 4 Defective Defective S-layer glycan biosynthesis initiation

enzyme WsfPfliB 1,263 270 6 WT Defective Lysine-N-methylase FliBdegS 1,161 268 1 Defective Defective Sensor protein DegSPBN151_2619 600 464 2 Defective Defective Type 11 methyltransferasePBN151_3232 2,034 141 4 Defective Defective Hypothetical proteinPBN151_3459 3,120 2,011 1 WT Defective Hypothetical extracellular proteinPBN151_4027 744 4 4 WT Defective ThiF family proteinPBN151_4312 1,086 768 4 WT Defective Capsular polysaccharide biosynthesis

proteins

Flagellar Gene Expression in Paenibacillus sp. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 3

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

strain exhibited normal motility (data not shown). In B. subtilis, although the phosphor-ylation site of DegU is essential for motility, degS is dispensable for motility (19, 20).Therefore, it is thought that a small amount of phospho-DegU (DegU�P), which isphosphorylated by small phospho-donors, such as acetylphosphate, is sufficient formotility in B. subtilis (19, 20, 32). A markerless degS deletion mutant was constructedand used to determine whether degS was required for motility in Paenibacillus sp.Motility was completely lost in the degS mutant (Fig. 1), and therefore, unlike in B.subtilis, DegS was indispensable for motility in Paenibacillus sp.

WsfP is an ortholog of Paenibacillus alvei WsfP, which is the enzyme that initiatesS-layer glycan biosynthesis (33, 34). Disruption of wsfP resulted in a complete loss ofmotility (Fig. 1). When the deleted wsfP region was restored using a pMAD derivativeplasmid carrying the wild-type wsfP sequence (see Materials and Methods), the result-ing strain exhibited normal motility (data not shown). These results indicate that cellsurface structures may be important for motility in Paenibacillus sp.

PBN151_4312 is a cytoplasmic protein of unknown function that is conserved inmany bacteria. PBN151_4312 contains the COG4421 domain of bacterial capsularpolysaccharide biosynthesis proteins (cell wall/membrane/envelope biogenesis). TheCOG4421 domain-containing protein WbqL is required for lipopolysaccharide synthesisin Caulobacter crescentus (35). Although its exact function is unclear, PBN151_4312 maybe involved in cell surface structures. The PBN151_4312 disruption mutant exhibited acomplete loss of motility on all media tested (Fig. 1). The entire region for PBN151_4312was cloned into a multicopy plasmid, pHY300PLK (36). Introduction of the resultantplasmid restored motility to the PBN151_4312 mutant, whereas the parental plasmidpHY300PLK did not restore it (data not shown).

FliB resembles lysine-N-methyltransferases, which are widely conserved in bacteria.FliB methylates Lys residues within flagellin in Salmonella enterica serovar Typhimuriumand Shewanella oneidensis, although the disruption of fliB has no effect on motility inthese bacteria (37, 38). Disruption of fliB resulted in a moderate defect in motility (Fig.1), and the fliB mutant was not able to spread on plates containing �1% agar.Complementation of fliB failed, as we were unable to clone the entire region of fliB inplasmid in E. coli.

FIG 1 Motility-defective mutants. Wild-type and mutant strains were inoculated onto the center of platescontaining 2� YT solidified with the indicated agar concentrations. Plates were incubated at 37°C for 18h. Plate diameter, 9 cm.

Kobayashi et al. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 4

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

PBN151_3459 is an unknown and probably extracellular protein that is conservedonly in Paenibacillus bacteria. The PBN151_3459 mutant was unable to spread only on1.5% agar (Fig. 1). The pHY300PLK derivative containing the entire region forPBN151_3459 restored motility to the PBN151_3459 mutant (data not shown).

Unlike the Tn10 insertion mutant, disruption of PBN151_0898 had no effect onmotility. We did not analyze PBN151_2619, PBN151_3232, and PBN151_4027 due tofailure of plasmid construction. Thus, among the genes identified by Tn10 mutagen-esis, we confirmed that at least five genes, degSU, wsfP, PBN151_4312, fliB, andPBN4312_3459, were required for motility.

Expression of flagellar genes in motility-defective mutants. Flagellin levels weretested in newly identified motility-defective mutants. The wild-type and mutant strainswere grown on medium containing 1.5% agar for 5 h, at which time the wild-type strainformed moving colonies. Cell surface-associated protein fractions were then preparedfrom cells, as described previously (28). A flagellar mutant of fliF, which encodes an Mring of the flagellar basal body, was used as a reference. Flagellin was strongly detectedas an abundant cell surface-associated protein in the wild-type strain and was notfound in the fliF mutant, as demonstrated by SDS-PAGE analysis (Fig. 2A). Mutants ofdegSU, wsfP, and PBN151_4312, which completely lost motility, exhibited no detectableflagellin. The fliB and PBN151_3459 mutants, which exhibited moderate defects inmotility, exhibited low levels of flagellin.

In addition to the loss of flagellin, the S-layer protein from the wsfP mutant migratedfaster than the S-layer protein from the wild-type strain (Fig. 2A). The deducedmolecular mass of the S-layer protein from its amino acid sequence is 111.5 kDa, whichmatched the size of the protein observed in the wsfP mutant; however, this was smallerthan the size (�150 kDa) observed in the wild-type strain. The S-layer protein isglycosylated in Paenibacillus alvei, where a homolog of WsfP catalyzes an initialreaction for S-layer glycan biosynthesis, and the deletion of wsfP results in a loss ofS-layer glycosylation in P. alvei (33, 34). These observations indicate that WsfP alsocatalyzes glycosylation of the S-layer protein in Paenibacillus sp. As describedearlier, PBN151_4312 contained the COG4421 domain of bacterial capsular poly-saccharide biosynthesis proteins, but the PBN151_4312 mutation had no effect onthe migration of the S-layer protein on SDS-PAGE (Fig. 2A).

Next, we examined transcription of flagellar genes in these mutants. Northern blotanalysis showed that transcription of flagellar genes hag, motAB, and motCD was

FIG 2 Flagellin levels of motility-defective mutants. (A) Wild-type and mutant strains were grown at 37°Cfor 5 h on 2� YT–1.5% agar medium. Cell surface-associated proteins were extracted, and the extractedproteins were separated by SDS-PAGE. S-layer and flagellin proteins are indicated by arrows. The S-layerprotein of the wsfP mutant is indicated by a dashed arrow. (B) The flgM mutation restores flagellin levelsto wsfP and PBN151_4312 mutants but not degSU, fliB, and PBN151_3459 mutants.

Flagellar Gene Expression in Paenibacillus sp. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 5

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

severely reduced in motility-defective degSU, wsfP, and PBN151_4312 mutants andmoderately reduced in the fliB and PBN151_3459 mutants (Fig. 3B). Thus, decreasedtranscription of flagellar genes resulted in reduced expression of flagellin in thesemutants.

We also found that flagellar gene transcription was moderately reduced in flagellinand flagellar stator deletion mutants of hag, motCD, and motAB motCD (Fig. 3B). Thisfinding suggests that motility defect itself may cause moderate reductions in flagellargene transcription. Thus, moderate reductions in flagellar gene transcription in fliB andPBN151_3459 mutations may be indirectly caused by their motility defects. Thereafter,we focused on degSU, wsfP, and PBN151_4312, which appeared to be more critical forflagellar gene transcription.

Promoter activity of flagellar genes. Transcription of flagellar genes is usuallyregulated by multiple promoters in bacteria (39). In B. subtilis, the early fla-che operonis mainly regulated by a �A-dependent promoter. Other “late” flagellar genes, includinghag and motAB, are regulated by �D-dependent promoters (reference 40 and refer-ences therein). �D is produced from the fla-che operon, but its activity is inhibited byan anti-�D protein, FlgM, until FlgM is excreted from the cytoplasm through thecompleted flagellar hook-basal body (41). The previous study showed that the com-position and organization of flagellar genes in Paenibacillus sp. are quite similar to thoseof B. subtilis (28). Homologs of �D and FlgM were conserved in the Paenibacillus sp. (28),and DNA sequences that resembled the consensus sequence of B. subtilis �D-dependent promoters were found upstream of the flagellar genes hag, motAB, andmotCD on the Paenibacillus sp. genome (Fig. 3A). Indeed, transcription of hag, motAB,and motCD greatly was reduced in the fliF deletion mutant, which did not form theflagellar hook-basal body (Fig. 3B). The flgM mutation restored transcription of these

FIG 3 (A) Putative �D-dependent promoters in Paenibacillus sp. Putative �10 and �35 regions areindicated by red characters. The consensus sequence for B. subtilis �D-dependent promoters is shown atthe bottom. (B) Transcription of flagellar genes in motility-defective mutants. Wild-type and mutantstrains were grown on 2� YT–1.5% agar plates. Transcription of hag, motAB, and motCD was analyzedby Northern blotting. rRNA stained with methylene blue is shown as a loading control.

Kobayashi et al. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 6

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

genes to the fliF mutant (Fig. 3B). These results indicate that the �D-FlgM regulationsystem is conserved in Paenibacillus sp.

To examine promoter activity, we cloned upstream regions of flgB (the first gene ofthe fla-che operon) and hag into the gfp reporter plasmid. The resulting PflgB-gfp andPhag-gfp transcriptional fusions were introduced into Paenibacillus sp. using the multi-copy plasmid pHY300PLK (36). Expression of green fluorescent protein (GFP) wasanalyzed microscopically. To test the system, GFP expression was assessed in thewild-type and fliF and flgM mutant strains. The fliF mutation had little or no effect onexpression of PflgB-gfp, but reduced expression of Phag-gfp, compared with that in thewild-type strain (Fig. 4). The flgM mutation had little or no effect on expression, but

FIG 4 Expression of PflgB-gfp and Phag-gfp transcriptional reporters. The indicated strains harboring PflgB-gfp orPhag-gfp reporters were grown on 2� YT–1.5% agar plates. Expression of GFP was analyzed by fluorescencemicroscopy. Phase-contrast and GFP fluorescence images are shown. Overlay images represent merged phase-contrast (false-colored red) and GFP (false-colored green) images.

Flagellar Gene Expression in Paenibacillus sp. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 7

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

expression of Phag-gfp was restored in the fliF fliM double mutant. These observationswere consistent with the Northern blot analysis (Fig. 3A), confirming the function-ality of the GFP reporter system. These results indicate that like in B. subtilis, flagellargene transcription is controlled by an early flagellar promoter for the fla-che operonand �D-dependent late flagellar promoters for hag and other flagellar genes inPaenibacillus sp.

Next, we examined the effect of the degSU, wsfP, and PBN151_4312 mutations onpromoter expression. The degSU mutation greatly reduced the expression of bothPflgB-gfp and Phag-gfp (Fig. 4). The flgM deletion restored the expression of Phag-gfp butnot the expression of PflgB-gfp to the degSU mutant (Fig. 4). Consistent with this, thedegSU flgM mutant did not express flagellin (Fig. 2B) and did not spread on agarmedium (Fig. 1). Similar to the degSU mutation, the degS mutation reduced expressionof PflgB-gfp and Phag-gfp (Fig. 4). The flgM mutation did not restore motility to the degSmutant (Fig. 1). Phosphorylated DegU directly activates the flgB promoter in B. subtilis(31). The requirement of degSU for flgB promoter activity indicates that phosphorylatedDegU probably directly activates the flgB promoter in Paenibacillus sp.

The wsfP mutation also reduced the expression of both PflgB-gfp and Phag-gfp;however, the effect on PflgB-gfp was weaker than with the degSU mutation (Fig. 4). TheflgM mutation fully restored the expression of both PflgB-gfp and Phag-gfp to the wsfPmutant (Fig. 4). Consistent with this, the wsfP flgM mutant expressed the wild-type levelof flagellin, as demonstrated by SDS-PAGE analysis (Fig. 2B). These results indicate thatthe wsfP mutation severely affects �D-dependent promoters rather than the flgBpromoter. Despite a healthy expression of flagellar genes, the wsfP flgM mutant stillexhibited a mild defect in motility on 1% and 1.5% agar (Fig. 1). This observationindicates that glycan chains on S-layer play important roles in Paenibacillus sp. motility.Since lipopolysaccharides serve as wetting agents facilitating swarming motility inProteus mirabilis and Salmonella enterica (42, 43), glycan chains may serve a similarfunction.

The PBN151_4312 mutation did not affect the expression of PflgB-gfp but reducedPhag-gfp expression (Fig. 4). The flgM mutation fully restored expression of Phag-gfp tothe PBN151_4312 mutant (Fig. 4). Consistently, the flgM PBN151_4312 mutant expressedthe wild-type level of flagellin and exhibited normal motility (Fig. 1 and 2B). Thus, thePBN151_4312 mutation affects only �D-dependent promoters.

As described above, the fliB and PBN151_3459 mutations did not have severe effectson flagellar gene expression. The flgM mutation did not restore flagellin levels ormotility to the fliB and PBN151_3459 mutants (Fig. 1 and 2B) and actually exacerbateda motility defect in the PBN151_3459 mutant (Fig. 1). These observations support ourspeculation that the fliB and PBN151_3459 mutations may not have direct effects onflagellar gene transcription.

DegSU is a regulator that responds to surface contact signals. Paenibacillus sp.induces flagellar gene transcription when grown on solid medium (28). wsfP andPBN151_4312 were probably involved in cell surface structures, which directly contactthe external environment. We investigated whether wsfP and PBN151_4312 wererequired to induce flagellar genes in response to surface growth conditions. Flagellargenes were not expressed in these mutants, but the wsfP and PBN151_4312 mutantsexpressed flagellin in the flgM mutant background (Fig. 2B). Therefore, flagellin levelswere compared in the flgM, wsfP flgM, and PBN151_4312 flgM mutants in liquid andsolid media. The flgM mutant exhibited higher expression of flagellin on solid mediumthan in liquid medium, as observed in the wild-type strain (Fig. 5A). Flagellin productionwas also induced on solid medium in the wsfP flgM and PBN151_4312 flgM mutants (Fig.5A). These results indicate that flgM, wsfP, and PBN151_4312 are not involved in theinduction of flagellar genes under surface growth conditions.

Next, we tested whether DegSU induced flagellar genes in response to surfacegrowth conditions. The degSU mutant did not express flagellin even in the flgM mutantbackground (Fig. 2A and B). Overexpression of two-component systems often causes

Kobayashi et al. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 8

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

signal-independent activation of these systems (44). If DegSU activates flagellar genetranscription in response to surface growth conditions, increased expression of DegSUmay bypass the requirement of signals for the induction of flagellar genes. To achieveoverexpression of degSU, a DNA fragment containing lacI and a LacI-repressive/isopropyl-thio-�-D-galactopyranoside (IPTG)-inducible spank promoter (45) was in-serted just upstream of the Shine-Dalgarno sequence for degS (Fig. 5C). The resultantPspank-degSU strain exhibited clear induction of degSU transcription in liquid and solidmedia in the presence of IPTG (Fig. 5D). Some transcription was also observed in theabsence of IPTG, probably because the spank promoter was leaky. In the absence ofIPTG, flagellin levels were low in liquid medium and increased on solid medium (Fig.5A). In contrast, in the presence of IPTG, high levels of flagellin were observed in bothliquid and solid media (Fig. 5A). Consistent with this observation, in the presence ofIPTG, the Pspank-degSU strain exhibited strong transcription of hag in both liquid and

FIG 5 DegSU activates flagellar expression in response to restriction of flagellar rotation. (A) Overex-pression of DegSU leads to constant flagellar expression regardless of growth medium. Wild-type andmutant strains were grown in 2� YT liquid medium (L lanes) or on 2� YT–1.5% agar solid medium (Slanes). The Pspank-degSU strain was grown in the absence or presence of 1 mM IPTG. Cell surface-associated proteins were analyzed by SDS-PAGE. S-layer and flagellin proteins are indicated by arrows.(B) Flagellin levels increased with increasing medium viscosity. The wild-type strain was grown to OD600

of 0.6 in liquid medium supplemented with indicated concentrations of Ficoll. Flagellin levels were thenanalyzed by SDS-PAGE. (C) The gene organization and transcriptional map of the degSU operon. Bentarrows indicate the position and direction of promoters. Dashed arrows of a, b, and c indicate transcriptsdetected by Northern blotting of degU in panel D. The deletion region of the degS mutant is indicatedbelow the gene map. The insert position of the lacI and spank promoter cassette in the Pspank-degSUstrain is also shown. (D) Northern blot analysis of degSU transcription in the Pspank-degSU strain.Wild-type and mutant strains were grown in 2� YT liquid medium (L lanes) or on 2� YT–1.5% agar solidmedium (S lanes). The Pspank-degSU strain was grown with or without 1 mM IPTG. Methylene blue-stainedrRNA is shown as a loading control. (E) The effect of overexpression of DegSU on flagellar expression.Wild-type and Pspank-degSU strains were grown in liquid medium (L), solid medium (S), or liquid mediumsupplemented with 15% Ficoll in the presence of 1 mM IPTG. hag transcription was analyzed by Northernblotting. rRNA stained with methylene blue is shown as a loading control. (F) The effect of flagellar statormutations on flagellar expression. Wild-type and ΔmotAB ΔmotCD strains were grown in liquid medium(L), solid medium (S), or liquid medium supplemented with 15% Ficoll.

Flagellar Gene Expression in Paenibacillus sp. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 9

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

solid media (Fig. 5E, left). These results indicate that overexpression of DegSU canbypass the requirement of signals to induce flagellar gene transcription. Thus, DegSUis a good candidate for a regulator that responds to surface growth conditions.

The activation signals for DegSU remain unknown, even in B. subtilis. However, asthe sensor kinase DegS is a cytoplasmic protein, the mechanism does not appear to bevia direct sensing of extracellular signals. Surface contact signals are sensed by flagellain Vibrio parahaemolyticus and Proteus mirabilis (4–9). Contact of flagellar filaments toa surface or viscous environment physically restricts their rotation, which triggersflagellar gene transcription (4–9). We hypothesized that restriction of flagellar rotationmight also induce flagellin expression in Paenibacillus sp. To test this hypothesis,flagellin levels were assessed in the wild-type strain cultivated in liquid mediumcontaining different concentrations of a viscous agent, Ficoll, which physically restrictsflagellar rotation. As shown in Fig. 5B, flagellin levels increased with increasing Ficollconcentrations. Specifically, the flagellin level was 3.2-fold higher in the presence of15% Ficoll than in the absence of Ficoll. Consistent with this, hag transcription wasclearly induced in the wild-type (WT) strain in the presence of 15% Ficoll (Fig. 5E, right).The effects of inhibition of flagellar rotation on flagellar transcription were furtherinvestigated in disruption mutants for flagellar stators, which were responsible forflagellar rotation. Paenibacillus sp. possesses two flagellar stator operons, motAB andmotCD, and motility was completely lost in the motAB motCD quadruple-deletionmutant (28). Transcription of hag was substantially elevated in the motAB motCDmutant in both liquid and solid media to levels comparable to those of the wild-typestrain grown on solid medium (Fig. 5F, left). The addition of Ficoll to liquid medium didnot further increase hag transcription in the motAB motCD mutant (Fig. 5F, right). Theseobservations indicate that Paenibacillus sp. also senses surface contact signals viarestriction of flagellar rotation. It is noted that, despite the high levels of hag transcrip-tion, flagellin production was largely lost in the motAB motCD mutant (Fig. 2A). Theflagellar stator proteins appear to be indispensable for flagellar assembly or flagellarstability in Paenibacillus sp.

The effect of the addition of Ficoll on hag transcription was also examined in thePspank-degSU strain in the presence of IPTG. The addition of Ficoll did not furtherincrease hag transcription (Fig. 5E, right). This observation indicates that restriction offlagellar rotation and DegSU may function in the same regulatory pathway. DegSU isshown to be activated by restriction of flagellar rotation in B. subtilis (17, 18). Takentogether, our observations indicate that DegSU may activate flagellar gene transcrip-tion in response to restriction of flagellar rotation in Paenibacillus sp.

Role for DegSU in B. subtilis motility. DegSU controls multiple cellular functions,such as motility, biofilm formation, poly-�-glutamate production, and protease produc-tion in B. subtilis (references 19 and 20 and references therein). Previous studies showedthat a small amount of DegU�P, which is probably phosphorylated by low-molecular-weight phospho-donors in a DegS-independent manner, is sufficient to induce flagellargene expression, whereas relatively a large amount of DegU�P, which is phosphory-lated by DegS, is required for biofilm formation, poly-�-glutamate synthesis, andprotease production (19, 20, 32). In B. subtilis, as in Paenibacillus sp., the disruption ofthe flagellar stator operon motAB or the addition of Ficoll to liquid medium induces thetranscription of DegSU-regulated genes (17, 18). However, as DegS is dispensable formotility and a large amount of DegU�P inhibits flagellar gene transcription in B. subtilis(19, 20, 46, 47), the activation of DegSU via restriction of flagellar rotation waspreviously proposed to lead to biofilm formation, poly-�-glutamate synthesis, andprotease production rather than increased motility (17, 18). The activation of DegSUthrough the restriction of flagellar rotation appears to constitute a part of the regula-tory network for biofilm formation. An exopolysaccharide synthesis enzyme, EpsE, is abifunctional protein, which binds to the flagellar rotor to inhibit flagellar rotation duringbiofilm formation (48). This action probably activates DegSU to enhance biofilmformation (21). However, since our findings suggest that the motility systems in the two

Kobayashi et al. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 10

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

bacteria share several common features, we speculated that the activation of DegSUthrough restriction of flagellar rotation might enhance motility in B. subtilis. In that case,the B. subtilis degS mutant would exhibit motility defects under certain conditions. Inprevious studies, motility of the degS mutant was tested at a single agar concentration(19, 20), and we therefore retested its motility on LB medium containing 0.3 to 0.9%agar. The wild-type strain (NCIB 3610) was able to spread on media up to 0.7% agar,whereas the degS mutant was able to spread on media up to 0.5% agar (Fig. 6). Underthe same conditions, the srfAC mutant, which was unable to produce surfactin, asurface-active molecule required for swarming motility but not swimming motility (49),spread on 0.3% agar medium but failed to spread on medium with �0.5% agar (Fig. 6).Thus, DegS was dispensable on 0.3% agar (swimming motility condition) and 0.5% agar(swarming motility condition), as described previously, but was indispensable on 0.7%agar medium (tougher swarming motility conditions). These results indicate that theactivation of DegSU can lead to increased motility under certain conditions in B. subtilis.

Conclusion remarks. Paenibacillus sp. forms moving colonies which can move on1.5% hard agar medium. The previous study showed that despite its unusual motilitybehavior, the composition and organization of flagellar genes are quite similar to thoseof B. subtilis (28). In this study, we showed that flagellar genes of Paenibacillus sp. werealso controlled by early and late promoters in a manner similar to that seen in B. subtilis.The early promoter was regulated by DegSU, as observed for B. subtilis. However, unlikein B. subtilis, DegS was essential for motility in Paenibacillus sp. This difference isprobably due to the function of SwrA. In B. subtilis, a small amount of DegU isphosphorylated by small phospho-donors independently of DegS (32). SwrA facilitatesbinding of a small amount of DegU�P to the promoter of the fla-che operon, whichinduces flagellar gene transcription in B. subtilis (14–16). However, since Paenibacillussp. possesses no SwrA homolog, DegS may be required to produce enough DegU�Pto bind and activate the fla-che operon promoter in Paenibacillus sp. Thus, induction offlagellar genes is mainly controlled by SwrA levels in B. subtilis (12), whereas it may becontrolled by DegU�P levels in Paenibacillus sp.

WsfP and PBN151_4312 were required for the activation of �D-dependent lateflagellar promoters. Since WsfP and PBN151_4312 functions were probably involved incell surface structures, �D activity may be subject to the regulation associated with cellsurface functions in Paenibacillus sp. Although similar regulations have not beenreported in B. subtilis, cell surface structures affect flagellar gene expression in Proteusmirabilis, in which the loss of lipopolysaccharides reduces flagellar gene transcription(50, 51). WsfP and PBN151_4312 were not required to induce flagellar genes inresponse to surface contact signal in the flgM mutant background. However, since cell

FIG 6 Motility of the B. subtilis degS mutant. B. subtilis strains were inoculated into the center of platescontaining LB solidified with the indicated agar concentrations. Plates were incubated at 37°C for 18 h.Plate diameter, 9 cm.

Flagellar Gene Expression in Paenibacillus sp. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 11

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

surface structures directly contact the external environment, the WsfP- andPBN151_4312-related regulations are expected to play important roles under surfacegrowth conditions.

Flagellar gene transcription was induced when Paenibacillus sp. was grown on solidmedium or in highly viscous liquid medium. Paenibacillus sp. cells sensed surfacecontact signals through restriction of flagellar rotation, as observed for the Gram-negative bacteria Vibrio parahaemolyticus and Proteus mirabilis (4–9). DegSU is knownto respond to restriction of flagellar rotation in B. subtilis (17, 18). Our observationsindicate that DegSU induces flagellar gene in response to restriction of flagellar rotationin Paenibacillus sp. Although the activation of DegSU through the restriction of flagellarrotation has been thought to enhance biofilm formation rather than motility in B.subtilis, we showed that the B. subtilis degS mutant exhibited motility defects under atougher swarming condition. This observation indicates that the activation of DegSUthrough restriction of flagellar rotation can enhance motility in both Paenibacillus sp.and B. subtilis, although this mechanism functions under limited conditions in B. subtilis.

The activation of DegSU plays indispensable roles in enhancing motility in Paeni-bacillus sp. In B. subtilis, DegSU regulates multiple cellular processes that occur underdifferent growth conditions (17–20, 52, 53), and its action and activity are regulated byaccessory proteins SwrA and DegQ, the latter of which is required for efficient phos-photransfer from phospho-DegS (DegS�P) to DegU (19). However, Paenibacillus sp.does not possess SwrA and DegQ homologs. The presence and absence of theseproteins would lead to differences in these bacteria.

MATERIALS AND METHODSStrains and growth conditions. The Paenibacillus sp. strains used in this study are shown in Table

2. Paenibacillus sp. was grown in LB (LB Lennox; Difco) or 2� YT (16 g · liter�1 tryptone [Difco], 10 g ·liter�1 yeast extract [Difco], 5 g · liter�1 NaCl) medium. Single colonies were isolated on 2.5% agar plates.Antibiotics were used at the following concentrations: erythromycin, 2.5 �g · ml�1; spectinomycin, 100�g · ml�1; chloramphenicol, 2.5 �g · ml�1; and tetracycline, 10 �g · ml�1. Ficoll 400 was obtained fromSigma-Aldrich. Transformation of Paenibacillus sp. was carried out using electroporation, as describedpreviously (28). The motility of Paenibacillus sp. was examined on 2� YT plates, as described previously(28). For protein and RNA analyses, overnight cultures were diluted 10-fold, and 100 �l was spread on9-cm-diameter plates. Cells were then cultivated at 37°C for 5 h prior to harvesting and analysis. B. subtilisstrain NCIB 3610 and the degS mutant were as described previously (19). Escherichia coli HB101 was usedfor plasmid construction.

Mini-Tn10 mutagenesis. Paenibacillus sp. was transformed with plasmid pIC333 (29). Single coloniesof fresh erythromycin-resistant transformants were inoculated into 10 ml of LB medium containingspectinomycin (LB-Spc) and grown at 28°C for 20 h with vigorous shaking. Cultures were then transferredto 1 liter of LB-Spc and incubated overnight at 40°C with vigorous shaking. The next day, 10 ml of eachculture was transferred to a new liter of LB-Spc and incubated overnight at 40°C with vigorous shaking.Part of each culture was centrifuged, and the cell pellet was suspended in the same volume of LBsupplemented with 20% glycerol and stored at �80°C. Each library was tested for its spectinomycin-resistant (Spcr) CFU and plated on LB-Spc–2.25% agar with an appropriate dilution, and the plates wereincubated at 37°C. Approximately 20,000 colonies from eight independent libraries were screened formutants that formed round colonies on 2.25% agar plates. Round colonies were picked, and theirmobility was subsequently tested on 1.5% agar in 24-well plates. Ninety motility-defective mutants wereisolated using this procedure. Chromosomal DNA was then isolated from the mutants as describedpreviously (28). Approximately 100 to 200 ng of chromosomal DNA was digested with TaqI (TaKaRa Bio)and then self-ligated overnight. The ligation mixtures were then precipitated with ethanol and used asthe template for PCR. DNA regions flanking the Tn10 insertions were amplified with the Tn10-F1/Tn10-R1or Tn10-F2/Tn10-R2 primer set (see Table S1 in the supplemental material). The resulting PCR productswere purified and used as the template for DNA sequencing.

Strain construction. Gene disruption was performed using plasmid pMAD, as described previously(28, 30). DNA fragments involved in gene disruption via double-crossover recombination were preparedby PCR and cloned into the pMAD plasmid (30). The primers used for PCR are listed in Table S1. Genedisruptions were obtained by a two-step screen (integration/excision) method, as described in supple-mental material. Plasmid pMAD was also used for complementation test. The deletion mutants weretransformed with pMAD containing the corresponding wild-type sequence, and complemented strainswere obtained by a two-step screen method. Plasmid pHY300PLK (36) was used as a vector forcomplementation and gfp reporters.

Fractionation of cellular proteins. Protein samples were prepared as described previously (28).Strains were grown to an optical density at 600 nm (OD600) of 0.6 at 37°C in 2� YT liquid medium orgrown for 5 h at 37°C on 2� YT–1.5% agar medium. Cells were suspended in 10 mM Tris-HCl (pH 7.6)and collected in 1.5-ml tubes. The number of cells in each sample was adjusted according to the OD600

Kobayashi et al. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 12

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

value. Cells were separated by centrifugation at 15,000 rpm for 2 min. The supernatant was used as thesecreted protein fraction. The precipitated cells were suspended in 10 mM Tris-HCl (pH 7.6)– 0.1% SDSand boiled for 1.5 min to extract cell surface-associated proteins. The samples were then centrifuged at15,000 rpm for 2 min, and the resulting supernatant was used as the cell surface-associated proteinfraction. The protein samples were then separated by SDS-PAGE (SuperSep Ace 10 to 20% gels; WakoPure Chemical Industries). The intensity of flagellin bands on SDS-PAGE was analyzed using the ImageJsoftware.

Northern blot analysis. Cells were collected from 2� YT liquid or plate cultures. Total RNA wasextracted, and Northern blot analysis was performed as described previously (19). The primers used forRNA probe preparation were described previously (28).

SUPPLEMENTAL MATERIAL

Supplemental material for this article may be found at https://doi.org/10.1128/AEM.00585-17.

SUPPLEMENTAL FILE 1, PDF file, 0.1 MB.

ACKNOWLEDGMENTThis work was supported by a Cooperative Research Grant from the Genome

Research for BioResource, NODAI Genome Research Center, Tokyo University of Agri-culture.

TABLE 2 Paenibacillus sp. strains used in this study

Strain Genotype or descriptiona Reference or strain construction

NAIST15-1 Prototroph 28P109 ΔdegSU::cat pMAD45¡NAIST15-1P151 ΔdegS pMADdegS¡NAIST15-1P116 ΔwsfP::cat pMAD88¡NAIST15-1P117 ΔPBN151_4312::cat pMAD96¡NAIST15-1P106 ΔfliB::cat pMAD60¡NAIST15-1P104 ΔPBN151_3459::cat pMAD25¡NAIST15-1P128 ΔflgM (in-frame deletion) pMADflgM¡NAIST15-1P154 ΔdegSU::cat ΔflgM (in-frame deletion) pMADflgM¡P109P268 ΔdegS ΔflgM (in-frame deletion) pMADflgM¡P1P152 ΔwsfP::cat ΔflgM (in-frame deletion) pMADflgM¡P116P160 ΔPBN151_4312::cat ΔflgM (in-frame deletion) pMADflgM¡P117P144 ΔfliB::cat ΔflgM (in-frame deletion) pMADflgM¡P106P145 ΔPBN151_3459::cat ΔflgM (in-frame deletion) pMADflgM¡P104P155 ΔfliF (in-frame deletion) 28P158 ΔmotCD 28P162 ΔmotAB ΔmotCD 28P101 Δhag::cat 28P159 ΔfliF (in-frame deletion) ΔflgM (in-frame deletion) pMADflgM¡P155P216 lacI Pspank-degSU pMADspankSU¡NAIST15-1P130 pHYG1-flgB (PflgB-gfp transcriptional fusion, Tetr) pHYG1-flgB¡NAIST15-1P170 ΔfliF (in-frame deletion) pHYG1-flgB pHYG1-flgB¡P155P166 ΔflgM (in-frame deletion) pHYG1-flgB pHYG1-flgB¡P128P171 ΔfliF (in-frame deletion) ΔflgM (in-frame deletion) pHYG1-flgB pHYG1-flgB¡P159P167 ΔdegS pHYG1-flgB pHYG1-flgB¡P151P136 ΔdegSU::cat pHYG1-flgB pHYG1-flgB¡P109P169 ΔdegSU::cat ΔflgM (in-frame deletion) pHYG1-flgB pHYG1-flgB¡P154P137 ΔwsfP::cat pHYG1-flgB pHYG1-flgB¡P116P168 ΔwsfP::cat ΔflgM (in-frame deletion) pHYG1-flgB pHYG1-flgB¡P152P143 ΔPBN151_4312::cat pHYG1-flgB pHYG1-flgB¡P117P172 ΔPBN151_4312::cat ΔflgM (in-frame deletion) pHYG1-flgB pHYG1-flgB¡P160P174 pHYG1-hag (Phag-gfp transcriptional fusion, Tetr) pHYG1-hag¡NAIST15-1P183 ΔfliF (in-frame deletion) pHYG1-hag pHYG1-hag¡P155P187 ΔflgM (in-frame deletion) pHYG1-hag pHYG1-hag¡P128P184 ΔfliF (in-frame deletion) ΔflgM (in-frame deletion) pHYG1-hag pHYG1-hag¡P159P180 ΔdegS pHYG1-hag pHYG1-hag¡P151P176 ΔdegSU::cat pHYG1-hag pHYG1-hag¡P109P182 ΔdegSU::cat ΔflgM (in-frame deletion) pHYG1-hag pHYG1-hag¡P154P177 ΔwsfP::cat pHYG1-hag pHYG1-hag¡P116P181 ΔwsfP::cat ΔflgM (in-frame deletion) pHYG1-hag pHYG1-hag¡P152P178 ΔPBN151_4312::cat pHYG1-hag pHYG1-hag¡P117P185 ΔPBN151_4312::cat ΔflgM (in-frame deletion) pHYG1-hag pHYG1-hag¡P160aTetr, tetracycline resistance.

Flagellar Gene Expression in Paenibacillus sp. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 13

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

REFERENCES1. Henrichsen J. 1972. Bacterial surface translocation: a survey and a clas-

sification. Bacteriol Rev 36:478 –503.2. Kearns DB. 2010. A field guide to bacterial swarming motility. Nat Rev

Microbiol 8:634 – 644. https://doi.org/10.1038/nrmicro2405.3. Partridge JD, Harshey RM. 2013. Swarming: flexible roaming plans. J

Bacteriol 195:909 –918. https://doi.org/10.1128/JB.02063-12.4. McCarter L, Hilmen M, Silverman M. 1988. Flagellar dynamometer con-

trols swarmer cell differentiation of V. parahaemolyticus. Cell 54:345–351.https://doi.org/10.1016/0092-8674(88)90197-3.

5. Stewart BJ, McCarter LL. 2003. Lateral flagellar gene system of Vibrioparahaemolyticus. J Bacteriol 185:4508 – 4018. https://doi.org/10.1128/JB.185.15.4508-4518.2003.

6. Gode-Potratz CJ, Kustusch RJ, Breheny PJ, Weiss DS, McCarter LL. 2011.Surface sensing in Vibrio parahaemolyticus triggers a programme ofgene expression that promotes colonization and virulence. Mol Micro-biol 79:240 –263. https://doi.org/10.1111/j.1365-2958.2010.07445.x.

7. Kawagishi I, Imagawa M, Imae Y, McCarter L, Homma M. 1996. Thesodium-driven polar flagellar motor of marine Vibrio as the mechano-sensor that regulates lateral flagellar expression. Mol Microbiol 20:693– 699. https://doi.org/10.1111/j.1365-2958.1996.tb02509.x.

8. Belas R, Suvanasuthi R. 2005. The ability of Proteus mirabilis to sensesurfaces and regulate virulence gene expression involves FliL, a flagellarbasal body protein. J Bacteriol 187:6789 – 6803. https://doi.org/10.1128/JB.187.19.6789-6803.2005.

9. Cusick K, Lee YY, Youchak B, Belas R. 2012. Perturbation of FliL interfereswith Proteus mirabilis swarmer cell gene expression and differentiation.J Bacteriol 194:437– 447. https://doi.org/10.1128/JB.05998-11.

10. Lele PP, Hosu BG, Berg HC. 2013. Dynamics of mechanosensing in thebacterial flagellar motor. Proc Natl Acad Sci U S A 110:11839 –11144.https://doi.org/10.1073/pnas.1305885110.

11. Tipping MJ, Delalez NJ, Lim R, Berry RM, Armitage JP. 2013. Load-dependent assembly of the bacterial flagellar motor. mBio 4(4):e00551-13. https://doi.org/10.1128/mBio.00551-13.

12. Mukherjee S, Bree AC, Liu J, Patrick JE, Chien P, Kearns DB. 2015.Adaptor-mediated Lon proteolysis restricts Bacillus subtilis hyperflagel-lation. Proc Natl Acad Sci U S A 112:250 –255. https://doi.org/10.1073/pnas.1417419112.

13. Kearns DB, Chu F, Rudner R, Losick R. 2004. Genes governing swarmingin Bacillus subtilis and evidence for a phase variation mechanism con-trolling surface motility. Mol Microbiol 52:357–369. https://doi.org/10.1111/j.1365-2958.2004.03996.x.

14. Calvio C, Celandroni F, Ghelardi E, Amati G, Salvetti S, Ceciliani F, GalizziA, Senesi S. 2005. Swarming differentiation and swimming motility inBacillus subtilis are controlled by swrA, a newly identified dicistronicoperon. J Bacteriol 187:5356 –5366. https://doi.org/10.1128/JB.187.15.5356-5366.2005.

15. Kearns DB, Losick R. 2005. Cell population heterogeneity during growthof Bacillus subtilis. Genes Dev 19:3083–3094. https://doi.org/10.1101/gad.1373905.

16. Mordini S, Osera C, Marini S, Scavone F, Bellazzi R, Galizzi A, Calvio C.2013. The role of SwrA, DegU and P(D3) in fla-che expression in B subtilis.PLoS One 8:e85065. https://doi.org/10.1371/journal.pone.0085065.

17. Cairns LS, Marlow VL, Bissett E, Ostrowski A, Stanley-Wall NR. 2013. Amechanical signal transmitted by the flagellum controls signalling inBacillus subtilis. Mol Microbiol 90:6 –21.

18. Chan JM, Guttenplan SB, Kearns DB. 2014. Defects in the flagellar motorincrease synthesis of poly-�-glutamate in Bacillus subtilis. J Bacteriol196:740 –753. https://doi.org/10.1128/JB.01217-13.

19. Kobayashi K. 2007. Gradual activation of the response regulator DegUcontrols serial expression of genes for flagellum formation and biofilmformation in Bacillus subtilis. Mol Microbiol 66:395– 409. https://doi.org/10.1111/j.1365-2958.2007.05923.x.

20. Verhamme DT, Kiley TB, Stanley-Wall NR. 2007. DegU co-ordinates mul-ticellular behaviour exhibited by Bacillus subtilis. Mol Microbiol 65:554 –568. https://doi.org/10.1111/j.1365-2958.2007.05810.x.

21. Belas R. 2013. When the swimming gets tough, the tough form a biofilm.Mol Microbiol 90:1–5. https://doi.org/10.1111/mmi.12354.

22. Ash C, Priest FG, Collins MD. 1993. Molecular identification of rRNAgroup 3 bacilli (Ash, Farrow, Wallbanks and Collins) using a PCR probetest. Proposal for the creation of a new genus Paenibacillus. Antonie VanLeeuwenhoek 64:253–260. https://doi.org/10.1007/BF00873085.

23. Genersch E, Forsgren E, Pentikäinen J, Ashiralieva A, Rauch S, Kilwinski J,Fries I. 2006. Reclassification of Paenibacillus larvae subsp. pulvifaciensand Paenibacillus larvae subsp. larvae as Paenibacillus larvae withoutsubspecies differentiation. Int J Syst Evol Microbiol 56:501–511. https://doi.org/10.1099/ijs.0.63928-0.

24. Lal S, Tabacchioni S. 2009. Ecology and biotechnological potential ofPaenibacillus polymyxa: a minireview. Indian J Microbiol 49:2–10. https://doi.org/10.1007/s12088-009-0008-y.

25. McSpadden Gardener BB. 2004. Ecology of Bacillus and Paenibacillus spp.in agricultural systems. Phytopathology 94:1252–1258. https://doi.org/10.1094/PHYTO.2004.94.11.1252.

26. Chou JH, Chou YJ, Lin KY, Sheu SY, Sheu DS, Arun AB, Young CC, ChenWM. 2007. Paenibacillus fonticola sp. nov., isolated from a warm spring.Int J Syst Evol Microbiol 57:1346 –1350. https://doi.org/10.1099/ijs.0.64872-0.

27. Roux V, Raoult D. 2004. Paenibacillus massiliensis sp. nov., Paenibacillussanguinis sp. nov. and Paenibacillus timonensis sp. nov, isolated fromblood cultures. Int J Syst Evol Microbiol 54:1049 –1054. https://doi.org/10.1099/ijs.0.02954-0.

28. Kobayashi K, Kanesaki Y, Yoshikawa H. 2016. Genetic analysis of collec-tive motility of Paenibacillus sp. NAIST15-1. PLoS Genet 12:e1006387.https://doi.org/10.1371/journal.pgen.1006387.

29. Steinmetz M, Richter R. 1994. Easy cloning of mini-Tn10 insertions fromthe Bacillus subtilis chromosome. J Bacteriol 176:1761–1763. https://doi.org/10.1128/jb.176.6.1761-1763.1994.

30. Arnaud M, Chastanet A, Débarbouillé M. 2004. New vector for efficientallelic replacement in naturally nontransformable, low-GC-content,Gram-positive bacteria. Appl Environ Microbiol 70:6887– 6891. https://doi.org/10.1128/AEM.70.11.6887-6891.2004.

31. Tsukahara K, Ogura M. 2008. Promoter selectivity of the Bacillus subtilisresponse regulator DegU, a positive regulator of the fla-che operon andsacB. BMC Microbiol 8:8. https://doi.org/10.1186/1471-2180-8-8.

32. Cairns LS, Martyn JE, Bromley K, Stanley-Wall NR. 2015. An alternateroute to phosphorylating DegU of Bacillus subtilis using acetyl phos-phate. BMC Microbiol 15:78. https://doi.org/10.1186/s12866-015-0410-z.

33. Zarschler K, Janesch B, Zayni S, Schäffer C, Messner P. 2009. Constructionof a gene knockout system for application in Paenibacillus alvei CCM2051T, exemplified by the S-layer glycan biosynthesis initiation enzymeWsfP. Appl Environ Microbiol 75:3077–3085. https://doi.org/10.1128/AEM.00087-09.

34. Zarschler K, Janesch B, Pabst M, Altmann F, Messner P, Schäffer C. 2010.Protein tyrosine O-glycosylation–a rather unexplored prokaryotic glyco-sylation system. Glycobiology 20:787–798. https://doi.org/10.1093/glycob/cwq035.

35. Cabeen MT, Murolo MA, Briegel A, Bui NK, Vollmer W, Ausmees N, JensenGJ, Jacobs-Wagner C. 2010. Mutations in the Lipopolysaccharide biosyn-thesis pathway interfere with crescentin-mediated cell curvature in Cau-lobacter crescentus. J Bacteriol 192:3368 –3378. https://doi.org/10.1128/JB.01371-09.

36. Ishiwa H, Shibahara H. 1985. New shuttle vectors for Escherichia coli andBacillus subtilis. II. Plasmid pHY300PLK, a multipurpose cloning vectorwith a polylinker, derived from pHY460. Jpn J Genet 60:235–243.

37. Frye J, Karlinsey JE, Felise HR, Marzolf B, Dowidar N, McClelland M,Hughes KT. 2006. Identification of new flagellar genes of Salmonellaenterica serovar Typhimurium. J Bacteriol 188:2233–2243. https://doi.org/10.1128/JB.188.6.2233-2243.2006.

38. Sun L, Jin M, Ding W, Yuan J, Kelly J, Gao H. 2013. Posttranslationalmodification of flagellin FlaB in Shewanella oneidensis. J Bacteriol 195:2550 –2561. https://doi.org/10.1128/JB.00015-13.

39. Chevance FF, Hughes KT. 2008. Coordinating assembly of a bacterialmacromolecular machine. Nat Rev Microbiol 6:455– 465. https://doi.org/10.1038/nrmicro1887.

40. Guttenplan SB, Shaw S, Kearns DB. 2013. The cell biology of peritrichousflagella in Bacillus subtilis. Mol Microbiol 87:211–229. https://doi.org/10.1111/mmi.12103.

41. Calvo RA, Kearns DB. 2015. FlgM is secreted by the flagellar exportapparatus in Bacillus subtilis. J Bacteriol 197:81–91. https://doi.org/10.1128/JB.02324-14.

42. Gygi D, Rahman MM, Lai HC, Carlson R, Guard-Petter J, Hughes C. 1995.A cell-surface polysaccharide that facilitates rapid population migration

Kobayashi et al. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 14

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from

by differentiated swarm cells of Proteus mirabilis. Mol Microbiol 17:1167–1175. https://doi.org/10.1111/j.1365-2958.1995.mmi_17061167.x.

43. Toguchi A, Siano M, Burkart M, Harshey RM. 2000. Genetics of swarmingmotility in Salmonella enterica serovar Typhimurium: critical role forlipopolysaccharide. J Bacteriol 182:6308 – 6321. https://doi.org/10.1128/JB.182.22.6308-6321.2000.

44. Stock JB, Ninfa AJ, Stock AM. 1989. Protein phosphorylation and regu-lation of adaptive responses in bacteria. Microbiol Rev 53:450 – 490.

45. Ben-Yehuda S, Rudner DZ, Losick R. 2003. RacA, a bacterial protein thatanchors chromosomes to the cell poles. Science 299:532–536. https://doi.org/10.1126/science.1079914.

46. Amati G, Bisicchia P, Galizzi A. 2004. DegU-P represses expression of themotility fla-che operon in Bacillus subtilis. J Bacteriol 186:6003– 6014.https://doi.org/10.1128/JB.186.18.6003-6014.2004.

47. Hsueh YH, Cozy LM, Sham LT, Calvo RA, Gutu AD, Winkler ME, Kearns DB.2011. DegU-phosphate activates expression of the anti-sigma factorFlgM in Bacillus subtilis. Mol Microbiol 81:1092–1108. https://doi.org/10.1111/j.1365-2958.2011.07755.x.

48. Blair KM, Turner L, Winkelman JT, Berg HC, Kearns DB. 2008. A molecular

clutch disables flagella in the Bacillus subtilis biofilm. Science 320:1636 –1638. https://doi.org/10.1126/science.1157877.

49. Kearns DB, Losick R. 2003. Swarming motility in undomesticated Bacillussubtilis. Mol Microbiol 49:581–590. https://doi.org/10.1046/j.1365-2958.2003.03584.x.

50. Morgenstein RM, Clemmer KM, Rather PN. 2010. Loss of the waaLO-antigen ligase prevents surface activation of the flagellar gene cas-cade in Proteus mirabilis. J Bacteriol 19:3213–3221. https://doi.org/10.1128/JB.00196-10.

51. Morgenstein RM, Rather PN. 2012. Role of the Umo proteins and the Rcsphosphorelay in the swarming motility of the wild type and anO-antigen (waaL) mutant of Proteus mirabilis. J Bacteriol 194:669 – 676.https://doi.org/10.1128/JB.06047-11.

52. Dartois V, Débarbouillé M, Kunst F, Rapoport G. 1998. Characterization ofa novel member of the DegS-DegU regulon affected by salt stress inBacillus subtilis. J Bacteriol 180:1855–1861.

53. Kobayashi K. 2015. Plant methyl salicylate induces defense responses inthe rhizobacterium Bacillus subtilis. Environ Microbiol 17:1365–1376.https://doi.org/10.1111/1462-2920.12613.

Flagellar Gene Expression in Paenibacillus sp. Applied and Environmental Microbiology

August 2017 Volume 83 Issue 15 e00585-17 aem.asm.org 15

on August 17, 2020 by guest

http://aem.asm

.org/D

ownloaded from